U.S. patent application number 10/743722 was filed with the patent office on 2005-07-07 for method and system for video encoding using a variable number of b frames.
Invention is credited to Dumitras, Adriana, Haskell, Barin Geoffry.
Application Number | 20050147167 10/743722 |
Document ID | / |
Family ID | 34710571 |
Filed Date | 2005-07-07 |
United States Patent
Application |
20050147167 |
Kind Code |
A1 |
Dumitras, Adriana ; et
al. |
July 7, 2005 |
Method and system for video encoding using a variable number of B
frames
Abstract
An effective method for dynamically selecting the number of I, P
and B frames during video coding is proposed. Short-term look-ahead
analysis of a video sequence yields a variable number of B frames
to be coded between any two stored pictures. The first picture of a
group of frames (GOF) may be coded as a B picture. Motion speed is
calculated for each picture of the GOF with respect to the first
picture of the GOF. Subject to exceptions, as long as the
subsequent pictures exhibit motion speeds that are similar and
motion vector displacements that are co-linear with those of the
first picture in the GOF, they may be coded as B pictures. When a
picture is encountered having a motion speed that is not the same
as that of the first picture in the GOF, the picture may be coded
as a P picture. In some embodiments, a sequence of B pictures that
terminates in a P picture may be called a "group of frames"
(GOF).
Inventors: |
Dumitras, Adriana;
(Sunnyvale, CA) ; Haskell, Barin Geoffry;
(Mountain View, CA) |
Correspondence
Address: |
KENYON & KENYON
1500 K STREET, N.W., SUITE 700
WASHINGTON
DC
20005
US
|
Family ID: |
34710571 |
Appl. No.: |
10/743722 |
Filed: |
December 24, 2003 |
Current U.S.
Class: |
375/240.16 ;
348/700; 375/240.12; 375/240.15; 375/240.24; 375/E7.148;
375/E7.151; 375/E7.164; 375/E7.165; 375/E7.181; 375/E7.211 |
Current CPC
Class: |
H04N 19/109 20141101;
H04N 19/172 20141101; H04N 19/139 20141101; H04N 19/87 20141101;
H04N 19/114 20141101; H04N 19/142 20141101 |
Class at
Publication: |
375/240.16 ;
375/240.15; 375/240.12; 348/700; 375/240.24 |
International
Class: |
H04N 007/12 |
Claims
We claim:
1. A method of assigning frame types for coding of pictures in a
video sequence, comprising: determining movement between image data
of a reference picture and each of a plurality of pictures in
temporal order, comparing the movement of a first picture in the
plurality, temporally closest to the reference picture, to movement
of other pictures therein, if the plurality of pictures exhibits
consistent motion speed, assigning the pictures therein as B
pictures.
2. The method of claim 1, further comprising assigning a picture in
the sequence that does not exhibit consistent motion speed with the
first picture as a P picture.
3. The method of claim 1, further comprising: determining whether a
scene change occurs in the plurality of pictures, and, if so:
coding a first picture temporally after the scene change as a P
picture, and coding all pictures in the plurality that occur from
the first picture to a picture immediately prior to the scene
change as B picture long as they exhibit consistent motion
speed.
4. The method of claim 3, further comprising coding the picture
before the scene cut as a P picture at full quality or low
quality.
5. The method of claim 1, further comprising coding a picture as a
P picture when the picture, if coded as a B picture, would cause a
number of consecutive B pictures to be larger than a predetermined
maximum.
6. The method of claim 1, further comprising coding select pictures
as I pictures pursuant to a random access policy.
7. The method of claim 6, further comprising coding a picture
temporally adjacent to and before the I picture as a P picture.
8. The method of claim 7, further comprising coding a picture
temporally adjacent to and before the I picture as a P picture
using full quality of low quality.
9. The method of claim 1, wherein the determining comprises
computing motion vectors between pixelblocks of the plurality of
pictures and of the reference picture and the comparing comprises
comparing motion vectors of the respective pictures.
10. A video coding method, comprising, from a sequence of video
data: determining a motion speed between a first picture and a
reference picture; for each picture following the first picture,
until a termination condition is met: determining a motion speed
for the respective picture, comparing the motion speed of a
respective picture with the motion speed of the first picture, and
coding the respective picture as a B picture if the motion speeds
are consistent with each other; and when the termination is met,
coding a picture as a P picture.
11. The video coding method of claim 10, wherein the termination
condition is met when motion speed of a picture is not consistent
with the motion speed of the first picture.
12. The video coding method of claim 10, wherein the termination
condition is met when a scene change is detected.
13. The video coding method of claim 12, further comprising coding
a frame immediately prior to the scene change in display order as a
P frame and coding a frame immediately after the scene change in
display order as an I frame.
14. The video coding method of claim 12, further comprising coding
frames immediately prior to the scene change in display order and
immediately after the scene change P frames.
15. The video coding method of claim 12, further comprising coding
frames immediately adjacent to the scene change in display order as
B frames within a group of frames.
16. The video coding method of claim 12, further comprising
detecting a scene change by comparing a correlation coefficient C
to a predetermined threshold, the correlation coefficient given by:
6 C ( n ) = i = 1 M j = 1 N x n ( i , j ) x n + 1 ( i , j ) i = 1 M
j = 1 N x n 2 ( i , j ) i = 1 M j = 1 N x n + 1 2 ( i , j ) ,
wheren and n+1 identify pictures between which the scene change may
be detected, x.sub.n(i,j) and x.sub.n+1(i,j) respectively represent
pixel values of pictures n and n+1 at pixel locations (i,j) and M
and N represent width and height (respectively) of pictures n and
n+1.
17. The video coding method of claim 10, wherein the termination
condition is met when a random access policy dictates that a
picture be coded as an I picture, and a picture temporally adjacent
to and before the I picture is coded as a P picture.
18. Apparatus, comprising: a memory to store pictures of a video
sequence, a video coder coupled to the memory, to code each stored
picture as one of an I picture, a P picture and a B picture, a
motion vector generator coupled to the memory to generate motion
vectors for a plurality of stored pictures, the motion vectors
measuring motion between a respective picture and a stored
reference picture, a colinearity detector having an input coupled
to the output of the motion vector generator, and a picture type
decision unit to assign pictures having generally consistent motion
speeds for B picture coding and to assign a picture that does not
exhibit consistent motion speed for coding as a P picture.
19. The apparatus of claim 18, wherein the picture type decision
unit further controls the video coder to cause it to code the B and
P pictures.
20. The apparatus of claim 18, wherein the video coder codes B and
P frames using motion vectors generated by the motion vector
generator.
21. The apparatus of claim 18, further comprising a scene change
detector coupled to the memory and to the picture type decision
unit, to identify stored frames that follow a scene change.
22. The apparatus of claim 21, wherein the picture type decision
unit assigns a frame immediately prior to the scene change in
display order for coding as a P frame and assigns a frame
immediately after the scene change in display order as an I
frame.
23. The apparatus of claim 21, wherein the picture type decision
unit assigns a frames immediately prior to the scene change in
display order and immediately after the scene change for coding as
P frames.
24. The apparatus of claim 21, wherein the picture type decision
unit assigns frames immediately adjacent to the scene change in
display order as B frames within a group of frames.
25. A coded video signal created from a method comprising: from a
sequence of video data, determining a motion speed between a first
picture and a reference picture; for each picture following the
first picture in the sequence, until a termination condition is
met: determining a motion speed for the respective picture, coding
the respective picture as a B picture if its motion speed is
consistent with the motion speed of the first picture; when the
termination is met, coding a picture as a P picture; and outputting
coded data of the B pictures and the P picture to a channel.
26. The coded video signal of claim 25, wherein the termination
condition is met when motion speeds of a following picture is not
constant with respect to the motion speed of the first picture.
27. The coded video signal of claim 25, wherein the termination
condition is met when a scene change is detected.
28. The coded video signal of claim 25, wherein the termination
condition is met when a random access policy dictates that a
picture be coded as an I picture, and a picture temporally adjacent
to and before the I picture is coded as a P picture.
29. A video coding assignment method, comprising, for each of a
plurality of pictures following a reference picture: adding a first
picture that follows the reference picture in display order to a
group of frames, determining a motion speed of the first picture
with respect to the reference picture, iteratively, for pictures
subsequent to the first picture in display order: adding the
subsequent picture to the group of frames, determining a motion
speed of the subsequent picture with respect to the reference
picture, if the motion speed of the subsequent picture is
consistent with the motion speed of the first picture, performing a
next iteration, and if not, coding the last picture of the group of
frames as a P picture and coding all other pictures in the group of
frames as a B picture.
30. The video coding assignment method of claim 29, wherein the
motion speed determinations are based on a pixelblock-by-pixelblock
comparison of motion vectors between the respective pictures.
31. The video coding assignment method of claim 30, wherein the
motion speed determinations are: 7 S ( n , b ) = d x ( n , b ) + d
y ( n , b ) n , whereS(n,b) represents the motion speed of a
picture, d.sub.x and d.sub.y represent displacements of the
respective pixelblock b of picture and n represents the temporal
distance of the picture from the reference picture.
32. The video coding assignment method of claim 29, wherein
colinearity is based on: 8 E ( n ) = 1 N blocks e ( n , b ) N
blocks , whereine(n,b) represents a difference of motion vector
displacements of the respective subsequent picture with respect to
the first picture, each scaled according to its temporal distance
from the reference picture, and N.sub.blocks represents the number
of pixelblocks in each picture.
Description
BACKGROUND
[0001] Video encoder optimization for bit rate reduction of the
compressed bitstreams and high visual quality preservation of the
decoded video sequences encompasses solutions such as
rate-distortion optimized mode decisions and parameter selections,
frame type selections, background modeling, quantization modeling,
perceptual modeling, analysis-based encoder control and rate
control.
[0002] Generally, many video coding algorithms first partition each
frame or video object plane (herein, "picture") into small subsets
of pixels, called "pixelblocks" herein. Then each pixelblock is
coded using some form of predictive coding method such as motion
compensation. Some video coding standards, e.g., ISO MPEG or ITU
H.264, use different types of predicted pixelblocks in their
coding. In one scenario, a pixelblock may be one of three types:
Intra (I) pixelblock that uses no information from other pictures
in its coding, Unidirectionally Predicted (P) pixelblock that uses
information from one preceding picture, and Bidirectionally
Predicted (B) pixelblock that uses information from one preceding
picture and one future picture.
[0003] Consider the case where all pixelblocks within a given
picture are coded according to the same type. Thus, the sequence of
pictures to be coded might be represented as
[0004] I1 B2 B3 B4 P5 B6 B7 B8 B9 P10 B11 P12 B13 I14 . . .
[0005] This is shown graphically in FIG. 5(a) where designations I,
P, B indicate the picture type and the number indicates the camera
or display order in the sequence. In this scenario, picture I1 uses
no information from other pictures in its coding. P5 uses
information from I1 in its coding. B2, B3, B4 all use information
from both I1 and P5 in their coding.
[0006] Since B pictures use information from future pictures, the
transmission order is usually different than the display order. For
the above sequence, transmission order might occur as follows:
[0007] I1 P5 B2 B3 B4 P10 B6 B7 B8 B9 P12 B11 I14 B13 . . .
[0008] This is shown graphically in FIG. 5(b).
[0009] Thus, when it comes time to decode B2 for example, the
decoder will have already received and stored the information in I1
and P5 necessary to decode B2, similarly B3 and B4.
[0010] The receiver then reorders the sequence for proper display.
In this operation I and P pictures are often referred to as "stored
pictures."
[0011] The coding of the P pictures typically utilizes Motion
Compensation predictive coding, wherein a Motion Vector is computed
for each pixelblock in the picture. Using the motion vector, a
prediction pixelblock can be formed by translation of pixels in the
aforementioned previous picture. The difference between the actual
pixelblock in the P picture and the prediction block, (the
residual) is then coded for transmission.
[0012] Each motion vector may also be transmitted via predictive
coding. That is, a prediction is formed using nearby motion vectors
that have already been sent, and then the difference between the
actual motion vector and the prediction is coded for transmission.
Each B pixelblock typically uses two motion vectors, one for the
aforementioned previous picture and one for the future picture.
From these motion vectors, two prediction pixelblocks are computed,
which are then averaged together to form the final prediction. As
above the difference between the actual pixelblock in the B picture
and the prediction block is then coded for transmission.
[0013] As with P pixelblocks, each motion vector of a B pixelblock
may be transmitted via predictive coding. That is, a prediction is
formed using nearby motion vectors that have already been
transmitted, and then the difference between the actual motion
vector and the prediction is coded for transmission.
[0014] However, with B pixelblocks the opportunity exists for
interpolating the motion vectors from those in the co-located or
nearby pixelblocks of the stored pictures. The interpolated value
may then be used as a prediction and the difference between the
actual motion vector and the prediction coded for transmission.
Such interpolation is carried out both at the coder and
decoder.
[0015] In some cases, the interpolated motion vector is good enough
to be used without any correction, in which case no motion vector
data need be sent. This is referred to as Direct Mode in H.263 and
H.264. This works particularly well when the camera is slowly
panning across a stationary background. In fact, the interpolation
may be good enough to be used as is, which means that no
differential information need be transmitted for these B pixelblock
motion vectors. Within each picture the pixelblocks may also be
coded in many ways. For example, a pixelblock may be divided into
smaller sub-blocks, with motion vectors computed and transmitted
for each sub-block. The shape of the sub-blocks may vary and need
not be square.
[0016] Within a P or B picture, some pixelblocks may be better
coded without using motion compensation, i.e., they would be coded
as Intra (I) pixelblocks. Within a B picture, some pixelblocks may
be better coded using unidirectional motion compensation, i.e.,
they would be coded as forward predicted or backward predicted
depending on whether a previous picture or a future picture is used
in the prediction.
[0017] Prior to transmission, the prediction error of a pixelblock
or sub-block is typically transformed by an orthogonal transform
such as the Discrete Cosine Transform or an approximation thereto.
The result of the transform operation is a set of transform
coefficients equal in number to the number of pixels in the
pixelblock or sub-block being transformed. At the receiver/decoder,
the received transform coefficients are inverse transformed to
recover the prediction error values to be used further in the
decoding.
[0018] Not all the transform coefficients need be transmitted for
acceptable video quality. Depending on the transmission bit rate
available more than half, sometimes much more than half, of the
transform coefficients may be deleted and not transmitted. At the
decoder their values are replaced by zeros prior to inverse
transform.
[0019] Also, prior to transmission the transform coefficients are
typically quantized and entropy coded. Quantization involves
representation of the transform coefficient values by a finite
subset of possible values, which reduces the accuracy of
transmission and often forces small values to zero, further
reducing the number of coefficients that are sent. In quantization
typically, each transform coefficient is divided by a quantizer
step size Q and rounded to the nearest integer. For example, the
transform coefficient C would be quantized to the value C.sub.q
according to: 1 C q = ( C + Q 2 ) Q
[0020] The integers are then entropy coded using variable
word-length codes such as Huffman codes or arithmetic codes.
[0021] The sub-block size and shape used for motion compensation
may not be the same as the sub-block size and shape used for the
transform. For example, 16.times.16, 16.times.8, 8.times.16 pixels
or smaller sizes are commonly used for motion compensation whereas
8.times.8 or 4.times.4 pixels are commonly used for transforms.
Indeed the motion compensation and transform sub-block sizes and
shapes may vary from pixelblock to pixelblock.
[0022] A video encoder must decide what is the best way amongst all
of the possible methods (or modes) to code each pixelblock. This is
known as the mode selection problem. Depending on the pixelblock
size and shape, there exist several modes for intra and inter
cases, respectively.
[0023] A video encoder must also decide how many B pictures, if
any, are to be coded between each I or P picture. This is known as
the frame type selection problem, and again, ad hoc solutions have
been used. Typically, if the motion in the scene is very irregular
or if there are frequent scene changes, then very few, if any, B
pictures should be coded. On the other hand, if there are long
periods of slow motion or camera pans, then coding many B-pictures
will result in a significantly lower overall bit rate. Moreover, a
higher number of coded B frames makes possible achieving
temporal/computational scalability at the decoder without impacting
greatly the visual quality of the decoded sequence and the
computational complexity of the decoder. Consequently, platforms
and systems with various CPU and memory capabilities can make use
of streams coded using numerous B frames.
[0024] Modern encoders typically select the number of B frames that
occur between each I or P picture to be equal to one or two. This
predetermined and somewhat arbitrary decision is motivated by
experimental work, which shows that for most video sequences the
above decision reduces the bit rate without affecting negatively
the visual quality of the decoded sequences. The opportunity
exists, however, to reduce the bit rate much more for sequences
that exhibit slow motion or camera pans by increasing the number of
B frames. It is believed that current coding systems do not take
advantage of this opportunity, due to (a) the difficulty of the
I/P/B decision and (b) the increase in the encoder's computational
complexity that the implementation of the frame type decision would
determine. Indeed, the appropriate number of B frames to be coded
for each sequence not only depends on both the temporal and spatial
characteristics of the sequence but it may vary across the sequence
as the motion characteristics often change and a selection of
different numbers of B frames for each different part of the
sequence is typically required. Accordingly, there is a need in the
art for a computationally inexpensive coding assignment scheme that
dynamically assigns a number of B pictures to occur between
reference pictures (I- and P- pictures) based on picture
content.
BRIEF DESCRIPTION OF THE DRAWINGS
[0025] FIG. 1 is a block diagram illustrating operation of a frame
type selector according to an embodiment of the present
invention.
[0026] FIG. 2 is a flow diagram illustrating a method according to
an embodiment of the present invention.
[0027] FIG. 3 is a graph illustrating ideal colinearity among the
motion vectors in a series of frames.
[0028] FIG. 4 illustrates operation of direct coding mode for B
frames.
[0029] FIG. 5 illustrates exemplary frame assignments in display
order and coding order.
DETAILED DESCRIPTION
[0030] Embodiments of the present invention provide a frame type
selector for a video coder. This selector assigns input pictures
from a video sequence for intra coding, predictive coding or
bidirectionally predictive coding. According to the embodiment, the
first picture following an I or P picture may be coded as a B
picture. For all pictures subsequent thereto, motion speed may be
calculated with respect to the reference picture, the I or P
picture. Subject to exceptions, as long as the subsequent pictures
exhibit generally similar, constant or almost constant motion
speed, they may be coded as B pictures. When a picture having an
irregular motion speed is encountered, then that picture may be
coded as a P picture. In some embodiments, a sequence of B pictures
that terminates in a P picture may be called a "group of frames"
(GOF). The frame with irregular motion speed may terminate a
current GOF.
[0031] FIG. 1 is a block diagram illustrating operation of a frame
type selector 100 according to an embodiment of the present
invention. The frame type selector 100 may include a picture buffer
110, a motion vector generator 120, a scene change detector 130, a
colinearity detector 140 and a picture type decision unit 150. The
picture buffer 110 stores video data of a current picture n and
furnishes it to the motion vector generator 120 and scene change
detector 130. A reference picture, a previous I or P picture, is
available to the motion vector generator 120 from storage in a
video coder 170. A previous picture n-1 (in display order) is
available to the scene change detector 130, provided by a buffer
pool 160.
[0032] The motion vector generator 120, as its name implies,
identifies relative motion between image information in the current
picture n and the reference picture of the previous GOF. Motion
vector calculation is well known in the video coding arts.
Generally, it involves comparing blocks of image data from a
candidate picture of video data (picture n) to blocks of image data
in the reference picture that are generally spatially co-incident.
If a matching block is found in the reference picture, the motion
vectors represent spatial displacement between the block's location
in picture n and the matching block's location in the reference
picture. Thus, a set of motion Vectors is generated for each
pixelblock in picture n. The motion vector generator 120 may output
motion vectors (labeled, "MV" in FIG. 1) to the colinearity
detector 140 and to the buffer pool 160. In the buffer pool 160,
the motion vectors of a picture n may be stored in association with
the video data for later use during video coding 170.
[0033] The colinearity detector 140 determines whether the motion
vectors of the new picture n demonstrate a general flow of motion
that is consistent with the flow of motion obtained from a prior
sequence of pictures (from the prior reference picture P to picture
n-1). The colinearity detector 140 may generate an output
representing a degree of difference between the colinearity of
motion vectors of picture n and the motion vectors of the first
picture in the GOF of the video sequence.
[0034] The scene change detector 130, as its name implies, can
identify scene changes in the source video data. Various scene
change detectors 130 are known in the art and can be integrated
into the system of FIG. 1. When a scene change is detected,
detector 130 indicates the change to the picture type decision unit
150.
[0035] The picture type decision unit 150 may determine how each
picture is to be coded. It generates control signals to the picture
buffer 160 and video coder 170 in response to these picture
assignment decisions. When the picture type decision unit 150
assigns the current picture n to be coded as a B-picture, it may
cause the video data of picture n and its associated motion vectors
to be stored in the buffer pool 160 to await later coding and may
advance operation to the next picture (picture n+1).
[0036] When the picture decision unit 150 determines that picture n
shall be coded as a P picture, the picture decision unit 150 may
enable the video coder 170, causing it to code all pictures of the
GOF stored in the buffer pool 160. All pictures that follow the
previously coded P picture, including the newly assigned P picture
and any B pictures that occur between the new P picture and the
previously coded P picture are coded by the video coder 170.
Operation of the frame type selector 150 may advance to a new input
picture n+1 and repeat the above decision steps using the frames of
the new GOF.
[0037] In an embodiment, the picture decision unit 150 also could
decide to code a picture as an I picture to satisfy other coding
policies that are provided to support random access to video frames
and the like. In this case, the picture type decision unit 150 may
also cause the video coder 170 to code all pictures resident in
buffer pool 160 up to and including the newly assigned I
picture.
[0038] As the foregoing description indicates, the frame type
selector 100 may process groups of frames from input video data.
Each GOF may have the form BB . . . BP (or, alternatively, BB . . .
BI). When input image data indicates generally consistent (i.e.,
similar, constant or almost-constant speed) motion among video
content, the pictures that exhibit the consistent motion are
assigned as B pictures to the extent possible. When the constant
motion speed terminates, a picture may be designated as a P
picture. The B pictures may be coded using the P picture of the
previous group of pictures and the newly identified P picture as
reference pictures. Because all of the B pictures are identified as
exhibiting generally constant motion speed, coding should be
particularly efficient.
[0039] FIG. 1 also provides a simplified block diagram of a video
coder 170. As explained above, the video coder 170 may include a
coding chain that generates residual pixel data from a comparison
of input video data and predicted video data (subtractor 180).
Residual pixel data may be subject to an orthogonal transformed
190, quantized 200 and entropy coded 210. Coding of elements
180-210 may be performed on each pixelblock of a picture. Coded
block data from the entropy coder 210 may be stored in a transmit
buffer 220, typically on a picture-by-picture basis, until it is
transmitted to a channel.
[0040] Video coders 170 typically include a decoding chain that
reconstructs image data in a manner that replicates operations to
be performed by a decoder (not shown) that receives coded video
data from a channel. Here, the decoding chain is shown as including
a decoder 230, a reference picture store 240, and a motion or
spatial predictor 250. The decoder 230 inverts operation of
elements 180-210 and generates reconstructed image data that can be
stored 240 as reference pictures for further prediction. Reference
pictures in storage 240 also may be input to the motion vector
generator 120 for use in building GOFs as described above. For
motion prediction in P and B frames, the motion predictor 250 may
forward selected image data from the reference pictures motion
vectors to the subtractor 180. For motion prediction in P or B
coding modes, the selected image data is identified by the motion
vectors, which In embodiments of the present invention, some of the
motion vectors can be generated by the motion vector generator
120.
[0041] According to an embodiment, the picture type assignment
techniques illustrated in FIG. 1 may be integrated into an overall
picture assignment policy that considers additional factors when
assigning coding types to individual pictures. In some instances,
for example, when applications require coding and transmission of I
frames at regular intervals to enable random access, a picture may
be coded as an I picture even if the frame type decision process of
FIG. 1 otherwise would assign the picture to P or B coding. Other
applications, such as videoconferencing applications, insert I
frames into a stream of coded video data at regular time intervals
to permit rapid synchronization if data were lost due to
transmission errors. Since an I frame has been coded without any
reference to other frames, decoding of the I frame would not be
affected by errors in prior frames.
[0042] FIG. 1 illustrates the picture buffer 110 and buffer pool
160 as discrete elements for purposes of illustration only. In
implementation, these elements may be provided as members of a
larger memory space for storage of video data generally.
[0043] In another embodiment, when a scene cut occurs between two
pictures n and n-1 and the picture before the scene cut n-1 is not
the first frame of GOF, then a picture type decision may assign
picture n-1 as a P frame and picture n as either an I or a P frame.
In this embodiment, the pictures n-1 and n may be coded at either
full quality of low quality. Full quality means using the same
coding parameters as for previous pictures. Low quality means
reducing the spatial quality of the picture, typically by
increasing the value of the quantization in 200.
[0044] In a further embodiment, when a scene cut occurs between two
pictures n and n-1, picture type decision may permit the GOF to
continue and assign a B frame to the next picture after the scene
cut (picture n). When the maximum number of B frames (decided in
the coding system) has been exceeded, a new frame may be assigned
for P coding, yielding a pattern PB . . . B.vertline..vertline.B .
. . BP (where .vertline..vertline. represents the position of the
scene cut). Optionally, B pictures that are members of a GOF that
includes a scene cut may be coded at low quality relative to
pictures from other GOFs.
[0045] The picture type decision scheme discussed so far provides
several advantages in video coding applications. First, because it
favors coding of consecutive pictures that exhibit similar motion
properties (i.e., constant or almost-constant motion speed) as B
pictures, it yields lower bit rates of the compressed streams.
Second, the picture type decision scheme is computationally
inexpensive. The computation of motion speeds and speed errors
requires simple operations. Moreover, the motion vectors computed
for the purpose of frame type decision are re-used during the
coding of B and P pictures. Thus, in the aggregate, the expense
associated with the picture type assignment scheme of the present
embodiments is minimal. Third, coding using several B pictures in
appropriate contexts also provides a simple form of scalability for
use with video decoders of varying capability. B pictures typically
are not reference pictures for other pictures and, therefore, some
video decoders can elect to drop selected B pictures to simplify
their decoding operation and still obtain useful reconstructed
data.
[0046] The picture type assignment scheme of the foregoing
embodiments provides advantages over, for example, a brute force
approach that simply would code every combination of B pictures and
pick the combination that minimized bit rate of the output coded
video signal. The brute force approach is far too complex. It would
requires a very large number of trial-and-error operations, most of
which must be discarded once a final decision is made. By contrast,
the present invention provides a frame type assignment scheme that
requires far less computational expense and higher efficiency--as
noted, motion vector computations from frame type assignment may be
re-used when the video data is coded.
[0047] FIG. 2 is a flow diagram illustrating a method according to
an embodiment of the present invention. The method 1000 may begin
with consideration of a new picture n from a sequence of video data
(box 1010). The method 1000 may determine if the new picture is the
first picture in the sequence (box 1020). If so, the method may
assign the picture's type as an I-picture and have the picture
coded (box 1030). Thereafter, the method 1000 may advance to the
next picture (box 1040) and return to box 1010.
[0048] For pictures other than the first picture in the video
sequence, the method 1000 may determine whether a scene cut has
occurred. In one embodiment, the method 1000 computes a correlation
coefficient between the current picture n and the previous picture
n-1 (box 1050). If the correlation coefficient is higher than some
predetermined threshold (box 1060), then the method 1000 may
determine that no scene cut occurred (box 1070). Thereafter, the
method may determine whether the nth picture causes a length of a
current group of pictures to meet a predetermined maximum length
set for the system (box 1080). If so, then picture n may be
assigned to be a P-picture (box 1090). The P-picture decision
terminates the current GOF (box 1100) and causes the video pictures
of the GOF to be coded (box 1110). Thereafter, unless the method
1000 has reached the end of the video sequence (box 1120), the
method advances to the next picture (box 1040) and repeats
operation (box 1010).
[0049] If at box 1080 the method 1000 determines that the n.sup.th
picture does not cause the maximum GOF length to be reached, the
method may compute forward motion vectors between picture n and the
reference picture of the previous GOF (typically, a P picture) (box
1130) and also compute the slope of the motion vector displacements
(box 1140). If the current picture n is the first picture of a new
GOF (box 1150), the method may assign the picture's type to be a
B-picture (box 1160) and advance operation to the next picture
(boxes 1040, 1010). Otherwise, the method 1000 may compute a speed
error from the displacement slopes of the current picture and the
first picture in the GOF (box 1170). If the speed error exceeds
some predetermined threshold (box 1180), then the picture may be
assigned as a P-picture (box 1090). Again, the P picture assignment
terminates a current GOF and causes pictures of the GOF to be coded
(boxes 1100, 1110).
[0050] If the speed error does not exceed the threshold (box 1180),
the method 1000 may determine whether the current picture is the
last picture of the video sequence (box 1190). If so, the method
1000 again may advance to box 1090, assign the picture to be a
P-picture and code the GOF (boxes 1100, 1110) before terminating.
Otherwise, the method 1000 may assign the current picture to be a
B-picture (box 1200) and advance operation to the next picture in
the video sequence (boxes 1040, 1010).
[0051] Returning to box 1060, if the correlation coefficient is
smaller than the scene cut threshold, the method 1000 may determine
that a scene cut occurred (box 1210). The method may assign a
picture type based on a scene management policy for the system (box
1220). In the simplest embodiment, the scene management policy may
dictate that the first picture following a scene cut shall be coded
as an I-picture. Other embodiments may assign to code the picture
as either an I-picture or P-picture depending upon the relative
bandwidth consumed by these coding choices. If the picture is
assigned to be an I-picture or a P-picture, the assignment
terminates the GOF (box 1100) and causes pictures therein to be
coded (box 1110).
[0052] Further, other embodiments may assign to code the picture
after the scene cut as the picture type decision dictates, with the
provision that, in the case such a decision is to encode the
picture as a B frame, measures are taken to prevent the B frame
from referencing any picture prior to the scene cut.
[0053] In one embodiment, a scene cut decision may be made based
upon a correlation coefficient established for each of two
temporally adjacent frames. A correlation coefficient C for a frame
n may be computed according to: 2 C ( n ) = i = 1 M j = 1 N x n ( i
, j ) x n + 1 ( i , j ) i = 1 M j = 1 N x n 2 ( i , j ) i = 1 M j =
1 N x n + 1 2 ( i , j ) , where
[0054] x.sub.n(i,j) and x.sub.n+1(i,j) respectively represent pixel
values at pixel locations (i,j) in pictures n and n+1, and M and N
represent the width and height of pictures n and n+1. By comparing
correlation coefficients for two adjacent pictures (e.g., pictures
n and n+1), scene changes may be detected. Small values of the
correlation coefficients imply that two adjacent pictures have
content that is sufficiently different to be classified as a scene
change. As noted, alternative scene change detection techniques are
permitted for use with embodiments of the present invention.
[0055] FIG. 3 is a graph illustrating ideal colinearity among a
series of pictures. As noted, motion vectors generally represent a
displacement between a block in a current picture and a closely
matching block from some reference picture. Displacement typically
is represented by two components, along the x and y axes.
Therefore, for a picture 1, a motion vector with the components
(d.sub.x1, d.sub.y1) may be obtained that measures the displacement
between picture 1 and a reference picture 0. Assuming a constant
time interval between pictures, colinearity would be observed in
picture 2 if the motion vector of the displaced block, having the
components (d.sub.x2, dy.sub.2) were twice the magnitude of the
motion vector for the block in picture 1. The block in picture 2 is
temporally displaced from the block in reference picture 0 twice as
much as the block in picture 1 and, therefore, the motion vectors
should be twice the size as those for picture 1 in conditions of
perfect colinearity. By extension, in conditions of perfect
colinearity, pictures 3, 4, 5 and 6 all should have motion vectors
that are equal to the motion vectors for picture 1 when scaled
according to the relative temporal displacements of each picture 3,
4, 5 and 6 to the reference picture 0. The motion vector components
d.sub.x, d.sub.y for each block and each picture would define lines
with a common slope as shown in FIG. 3.
[0056] In practice, of course, perfect colinearity will not always
be observed. Accordingly, the motion vector of the first picture in
a GOF (picture 1 in the example of FIG. 3), may be selected as the
reference with respect to which the speed errors (i.e., the slope
errors) are computed. Successive pictures may be tested to
determine whether the slopes of motion vector displacements for
those pictures are within suitable tolerances of the reference
slope and, if so, to include the pictures in a GOF as B pictures.
When a picture's displacement slope falls outside the defined
tolerances, the GOF may be terminated.
[0057] According to an embodiment, motion vectors may be determined
for all pixelblocks in a candidate picture. Again, let d.sub.x and
d.sub.y by the components of a motion vector (displacements) along
the x and y directions. If a scene cut does not exist between a
first picture of a GOF and the preceding picture, it can be assumed
that the first picture of the GOF is a B-picture (picture no. 1).
Starting with the first picture (picture 1), for each picture of
the GOF, the system may compute the motion speed. The motion speed
of a block b in the picture may be measured by slopes S.sub.x(n, b)
and S.sub.y(n, b) and S(n, b) as follows: 3 S x ( n , b ) = d x ( n
, b ) n ( 1. ) S y ( n , b ) = d y ( n , b ) n ( 2. ) S ( n , b ) =
S x + y ( n , b ) = d x ( n , b ) + d y ( n , b ) n ( 3. )
[0058] Starting with picture 2, motion speed error may be
calculated with respect to the motion speed of the first picture
(B1) of the GOF:
e.sub.x(n, b)=S.sub.x(n, b)-S.sub.x(1, b) (4.)
e.sub.y(n, b)=S.sub.y(n, b)-S.sub.y(1, b) (5.)
e(n, b)=e.sub.x+y(n, b)=S(n, b)-S(1, b) (6.)
[0059] Thus, an error value can be obtained for each image block in
the candidate picture. The system may compute the speed error for
picture n (i.e., E(n)) as the mean of absolute speed errors of all
blocks in the picture, in which case E(n) is given by: 4 E ( n ) =
1 N blocks e ( n , b ) N blocks ( 7. )
[0060] where N.sub.blocks represents the number of pixelblocks per
picture. As long as the error of a picture is less than a
predetermined threshold value, that picture may be added to a group
of pictures as a B picture. If not, then the picture may be coded
as a P or I picture and the current group of pictures may be
terminated.
[0061] The foregoing picture type decision scheme contributes to
highly efficient coding of pictures. At a high level, the picture
assignment scheme identifies pictures that exhibit a common motion
speed and small speed errors among them. When these characteristics
are identified, the picture type decision scheme classifies a
relatively large number of candidate pictures as B pictures.
[0062] Tying frame type decisions to an observable pattern of
motion speeds among pictures also can yield additional advantages
in terms of coding effectiveness. Pictures may be coded according
to the direct mode in H.263+ and H.264. In other words, not only
more B frames are encoded, but they are also coded effectively. As
illustrated in FIG. 4, in temporal direct mode motion vectors are
interpolated from those in the co-located pixelblocks of the stored
pictures. For a pixelblock in a B picture, the co-located
pixelblock is defined as a pixelblock that resides in the same
geometric location of the first reference picture in list 1, where
list 1 and list 0 are lists of reference pictures stored in a
decoded picture buffer. Given the motion vector of the co-located
block mv.sub.col, motion vectors mv.sub.L0 and mv.sub.L1 can be are
interpolated with respect to the reference pictures in lists 0 and
1 (respectively) according to: 5 mv L0 ( n ) = k T b T d mv col (
8. ) mv L1 ( n ) = k T d - T b T d mv col ( 9. )
[0063] where n represents the picture for which the frame type
decision is being made, and k is a constant that includes a
distance scale factor and rounding. All motion vectors have x and y
components. Notations T.sub.b, T.sub.d represent differences
between the picture order counts according to:
T.sub.b=DPOC(F.sub.n, F.sub.L0)
T.sub.d=DPOC(F.sub.L1, F.sub.L0)
[0064] where F.sub.n, F.sub.L0, F.sub.L1 denote the current frame,
a reference frame from list 0 and a reference frame from list 1,
respectively. Of course, direct mode interpolation may be performed
for all B pictures in a GOF such as those shown in phantom in FIG.
4.
[0065] Several embodiments of the present invention are
specifically illustrated and described herein. However, it will be
appreciated that modifications and variations of the present
invention are covered by the above teachings and within the purview
of the appended claims without departing from the spirit and
intended scope of the invention.
* * * * *